bug#5618: 23.1.91; shell-command-on-region unexpectedly erases buffer

[tomas@tuxteam.de]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thanks, Glenn for looking into it.

On Tue, Feb 23, 2010 at 01:08:37PM -0500, Glenn Morris wrote:
> tomas@tuxteam.de wrote:
> 
> > The whole content of the buffer gets replaced by output of
> > command. According to the doc, I'd expect the ouput to be inserted
> > before point.
> 
> I guess this is a documentation bug, in the same way as was
> 
> http://debbugs.gnu.org/4891

Yes, I think it's the same "bug". Now if I knew how to merge that...

> I'm not sure that the installed doc fix for that bug is actually
> correct. The suggestion in
> 
> http://debbugs.gnu.org/4891#10

Except that in our case, even if it's the current buffer, the whole
content of the buffer is overwritten as well, rendering the above
mentioned doc fix incorrect:

    "If OUTPUT-BUFFER is not a buffer and not nil,
       insert output in current buffer.  (This cannot be done
       asynchronously.)  In the latter case, the output is inserted
       after point (leaving mark after it)"

Besides, the behaviour in the current documentation seems way more
useful (i.e. _not_ erasing the buffer whenever REPLACE is non-nil).

What do you think?

> seems correct to me.

Half of it, yes, see above.

Regards
- -- tomás
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFLhMAFBcgs9XrR2kYRAghcAJ97B90JiuOcu8Ck+Yt0iCPtRQcw1gCePIY8
XYLgohnxDDQjrsn1OPJf6mY=
=4Z+h
-----END PGP SIGNATURE-----